GIFGIF—Measuring Emotional Responses To GIFs

By Joelle Renstrom | 7 years ago

(these GIFs move on their site)

Anyone who spends time on the internet is familiar with GIFs, which are the perfect way to procrastinate online. Shorter than a video but more dynamic than a picture, GIFs encapsulate popular culture in a more immediate and accessible way than perhaps any other nugget from the internet. And, in case you’re wondering, they’re pronounced JIF, not GIF, though they’ll always be GIFs to me — that is, unless they’re GIFGIFs. That’s right, now you can have double the GIFs for twice the flavor, twice the fun.


Travis Rich and Kevin Hu, a pair of MIT Media Lab students, realized that GIFs have the potential to be more than just a tiny representation of a moment or idea. They have can convey emotion — and in a far more real way than emoticons ever could. They devised the GIFGIF project as a way to measure GIFs’ emotional expressiveness, allowing people to search and access GIFs via emotional conveyance, rather than just subject matter. GIFs vary a lot — they’re often goofy, but they’ll also represent a moment that seems to perfectly capture an emotion, such as anger, grief, fear, desire, etc. How can GIFs achieve that effect? Rich and Hu wondered about the different ways people can interpret GIFs, especially depending on cultural background. So they’ve set up a site where people can vote for the best GIF expression of a particular emotion.

Right now, their index contains 6,118 GIFs. They’re tallying up the votes and adding new GIFs once enough data comes back on the ones they already have. When a user viewers two GIFs side by side and picks the one that best expresses the emotion, Rich and Hu realize that there’s more than just a single emotion being conveyed, though they try to hone in on the most basic one. The six most universal emotions, according to psychologist Paul Ekman, are anger, disgust, fear, happiness, sadness, and surprise. Ekman later added 11 more core emotions, and those are the ones incorporated into GIFGIF.

Rich and Hu, who were inspired by their MIT colleagues’ project Place Pulse, which asks users to pick the “livelier” place based on a pair of photographs, believe that GIFs are instrumental when it comes to non-verbal communication, and could be used much like emoticons for GIF-messaging. GIFGIFs could be used in psychology to measure the interpretation of emotions, especially when it comes to mental health and cultural backgrounds, as well as in the classroom. Thus far, they’ve noticed trends in the data they’ve collected, especially concerning the location of the voters, but they aren’t quite ready to release the numbers yet.

The project has implications for the future of AI and human interaction with computers and robots as well. Instead of error messages, computers could offer up GIFs of the related emotion. GIFs could also self-correct depending on the user’s interpretation of various emotions.